• State
  • Locked Locked
  • Replies 3 replies
  • Subscribers 46 subscribers
  • Views 225 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AFE5818: unreproducable dead zone in AFE5818 model

Jayson Vivet
Prodigy 20 points
Part Number: AFE5818
Other Parts Discussed in Thread: AFE5808

Presentation for the problem for TI 

 

We encounter a problem when sending a long burst of pulses on the AFE5818. 

The problem is also the same with AFE5808. 

 
 

In the drawing above, we see the signals at the input of the System : 

 in blue : INP1, in Red : INM1 in Purple : difference of both. 

The problem we encounter is due to the fact that the polarization offsets moves on the positive input and it takes time for INM and INP to go back to their initial level: during this time, we find dead zone (unusable for measurement) due to the amplification of the signal through LNA, VCAT, PGA. Even though there are Integrators/HPF in the LNA, PGA and ADC, the signal in the dead zone is not usable. 

In one of our design, we decided to connect the ACT Pin and, in another design, ACT isn’t connected (floating input).  

We tried to use the simulation model of the AFE5808 available on TI’s website, to reproduce the problem to tests different solutions  

On channel 1 we tried the version recommended on the datasheets  

On channel 4 the design as we implemented on one of our boards (I added a 100Meg Resistance to simulate a NC connection else the simulation doesn’t work) 

 

Trying to simulate the problem 

On channel 1 :  

 

OnChannel 4 : 

 

The simulation gives no difference with and without ACT capacitor, and in addition, the moving bias effect is not seen.  

Is there something missing in my simulation? Or does the model not include this phenomenon? If so, could someone in TI fix the issue? 

If the issue can’t be fixed, perhaps you’ve encountered this issue in the past? Is there a record of a possible solution?  

Kind regards. 

  • 0
    TI__Expert 4661 points

    Hi,

    It seems you are applying 2Vpp signal. It is large and will cause discharge of INP cap since input pair of AFE is BJT. In simulation model such leakage current is not modeled hence not showing in simulation. However this is usual phenomenon. Device will correct the input discharge using LNA HPF response. So if you want to see fast recovery, you can try increasing the LNA HPF corner frequency. 

    Thanks!

    Regards,

    Shabbir

  • 0 in reply to Shabbir Amjhera Wala
    Prodigy 20 points

    Hello  and kind thanks for answering our posts.

    We've performed further tests using the LNA integrator and confirmed what you're saying. However, modifying Cbypass (15 nF) to a lower value induces oscillations and therefore doesn't reduce the dead zone.

    To help us fix our issue, we'd like to be able to simulate the issue and help us undertand the phenomenon. By any chance can someone in TI produce a new model that includes this dead zone?

  • 0 in reply to Jayson Vivet
    TI__Expert 4661 points

    Hi,

    Modelling the overload condition is tricky. 

    I have one more suggestion. See if it is helpful:

    - First use the LNA gain of 12dB. Lower LNA gain allows higher input swing support.

    - You can set reduce the input clamp level. Try these two settings:

    - Use 200kHz LNA HPF corner by programming LNA_HPF_PROG register.

    Thanks!

    Regards,

    Shabbir